The realization of structure based molecular design is linked to our ability to accurately estimate or predict protein stabilities or binding affinities from structural considerations. This task essentially reduces to the ability of predicting the Gibbs energy that accounts for the stability of a protein or the binding affinity of a complex from a detailed knowledge of their structure. Recently, it has been realized that the thermodynamic parameters (deltaCp, deltaH, deltaS) that define the Gibbs energy of protein stabilization or binding can be parametrized in structural terms. This empirical approach appears to be promising and to have predictive potential, provided that the structural parametrization is extended and refined. The main goal of this proposal is precisely to extend and refine the existing structural parametrization of the thermodynamic parameters deltaCp, deltaH, and deltaS using systems in which the effects of specific contributions have a higher impact and are therefore easier to identify and candidate. These systems include folding/unfolding of specific mutants of low molecular weight proteins, and the binding of different peptides to specific protein binding sites. The thermodynamic parameters for those events will be measured directly by high sensitivity calorimetric techniques. In our studies, we will consider specific mutants of a 33 amino acid peptide corresponding to the coiled coil (leucine zipper) region of GCN4, and the binding of specific inhibitors to aspartic proteases. The protease systems include endothiapepsin for which the structure of several inhibitor complexes is known at high resolution and the human immunodeficiency virus (HIV) protease for which high resolution structural information is also available. Besides providing information of general significance and applicability, the systems chosen are of intrinsic importance and medical relevance.